Octavia Experiments/Improvements

5 03 2009

I finally got around to trying the Octavia circuit. I’ve been fooling with it on breadboard in my spare time the past couple days, and I’ve figured out some interesting things.
Just to make sure we’re on the same page, I’m using the Tycobrahe Octavia schematic on Analogguru’s site.
http://analogguru.100free.com/193/schematics/Tycobrahe_Octavia.gif
There are a couple minor differences in my test circuit; firstly, I’ve oriented the MPSA18 the correct direction with the collector towards earth. I did try it the original backwards way, and it does work, but it sounds so bad it’s not worth messing with to try and be “vintage correct” with Tycobrahe original. For the other two transistors, I’m using 2N2907… not for any special reason, that’s just what suitable PNP silicon transistors I had handy. They do the job fine. The transformer is the Mouser 42TM022 that Small Bear sells for DIY Octavia. I’m using 1N34A for the diodes. Most of the resistor values are the same as the schematic with a couple little subs where I didn’t have the exact value, but nothing that would make a difference.

So, I build up the basic “stock” circuit. It works, but it sounds rather… ugh… not so great. It gets octave, but it’s only really happening when you play high up on the neck. I know that’s a well-known characteristic of it, but that doesn’t mean it has to be that way. It also gets incredibly splattery/mushy when you turn the fuzz up. I guess this is normal too. The problem is that the distortion of the fundamental totally hides the octave when the fuzz is up… you can hear it better with lower fuzz settings, but it’s still very dependent on dynamics and working the strings to get the note to come out. I suppose this is a testament to Hendrix’s mastery of the effect.

I listen to these things, and to me, they sound like problems that can be solved. So, I play and listen… and what do my ears tell me? They tell me there is too much low frequencies saturating the primary of the transformer and that the overall response is too slow to dynamics. I try a much smaller capacitor coupling from the third transistor to the transformer – C6 in AG’s schematic – this is normally 33uF… I try 330nF…..
Now there is octave all the way up and down the neck; on the low strings, too! Then, I replace C7 (220uF) with 47uF; dynamics are now much improved; less slow compression at high fuzz settings.

I don’t like the diodes going straight to the volume control and then the output; it looks like you would get inconsistent results with different amps/effects after it, and that might affect how well it does the octave thing. I put a buffer after the diodes with another PNP transistor. Works great. Put the volume control at the very end after the buffer. The signal coming from the diodes is low enough that the buffer stays clean even with the full output going to it.

Now, I want to be able to have both the “traditional” Octavia splat sound as well as my new improved octave, so I did a compromise and tried a 10uF coupling cap with a 100k pot to fade it in on top of my new lower value C6. Works very well at dialing in two very different responses; though you may want to experiment with tapers/values for the pot. I keep my new value of 47uF for C7 in any case because it sounds better than 220uF regardless.

One more thing I tried: put a resistor in series with the transformer’s primary connection to ground. The idea here is to reduce the loading of the primary on the stage that’s driving it. This is another attempt at reducing the amount of squishy-ness in the sound. I just tried a 1K resistor. It worked to a degree. As expected, the output volume dropped a bit because the driving signal is now being divided between the primary and the resistor, but there was still enough output to be usable. This gave a very dramatic change to the timbre of the octave effect… very hard to describe; it was still a fuzzy octave sound, but just sounded different than it did before. To me, it sounded like the octave wasn’t as pronounced in a subtle way, so I dropped the resistor, but it has given me reason to experiment with a transformer that has a higher-impedance primary in the future.

My conclusion is that the transformer and how it interacts with the drive circuit is absolutely the determining factor in the character of sound you get from this. The changes I made to try and improve it were all centered around changing how the circuit responds with the transformer, so I’m certainly going to get different transformers to play with.

I’m going to build one up based on what I have on the breadboard now – it sounds really neat and I was able to get some super-cool sounds with different pickup selections and different playing techniques. I totally got this cool “Robot Blues” thing happening.

I’ll draw up a schematic and share the final design as a full DIY project once I get it put together. But, before that, I have a bunch of Four Bangers that need built!

-MC

Advertisements

Actions

Information

2 responses

20 05 2009
mitchell

Nice article!

There was a suggestion on DIYStompoboxes to apply a voltage to the center tap to bias the output voltage of the coil up to the diode voltage. It looks like the Bobtavia might make use of this.

I built a Green Ringer some time back using the Filter mod posted at GGG. This really accentuates the octave effect to an almost cartoonish extreme. It’s my guess that this could be applied to any octave circuit before the rectifier stage.

21 05 2009
Martin

That reminds me of the diode biasing arrangement you can see in the Foxx Tone Machine and Gus’ Octave Up Sick Box. Definitely something to try – thanks!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s




%d bloggers like this: